Present day approaches to uncover gene function often rely on the ability to construct gene knockouts. While this gene-by-gene approach is straightforward in organisms with facile genetic methods such as yeast and flies, it is difficult in most other organisms. Furthermore, gene knockouts only allow analysis of nonessential genes. Fast acting, reversible chemical probes are an ideal way to perturb the function of essential genes. The primary objective of this proposal is to identify inhibitory chemical probes for all essential gene products in yeast using a genome-wide method that allows us to profile the relative sensitivity all protein targets in a cell simultaneously. This chemogenomic assay, Haplolnsufficiency Profiling (HIP), is based on the observation that lowering the dosage of a single gene from two copies to one copy in diploid yeast results in a heterozygote that is sensitized to any compound that acts on the product of this gene (Nat Genet. 21,278 (1999)). In the assay, a complete collection of 5,918 heterozygous deletion strains is pooled, grown competitively in the presence of compound and sampled. Molecular bar-codes incorporated into each strain allow parallel analysis and relative strain fitness to be quantitatively assessed by hybridization to oligonucleotide arrays. We validated the assay by screening 1,200 compounds with known biological activity. For all FDA therapeutics screened, the strain heterozygous for the known target was highly sensitive to drug. Recent breakthroughs in assay miniaturization and automation enable high-throughput screening of large numbers of compounds. As a result, the assay can now be implemented in a combined functional genomics and compound discovery effort. The outcome of this effort will be a catalog of chemical probes and their essential gene targets. In addition, such genome-wide exploration of biological and chemical space will identify novel, "druggable" targets and allow characterization of chemical structures that comprise "drug-like" compounds. As part of this effort, data will be disseminated to the community via a web interface. [unreadable] [unreadable]